[rules-users] BRMS rule info logging

techluver techluver007 at gmail.com
Mon Oct 6 13:29:51 EDT 2008


David or someone please respond tome.

1.I would fist step back and ask the question if a rules engine is the right
solution to your problem.

- Actually I do have same question and need experts help to found out right
answer.

1.Is this data you are getting in daily in a single table, multiple tables?
- we want to assume that it is from multiple tables.

2.How complex is the data? Is the data simple enough that rules could just
be
some SQL statements/Stored procs?

- data may not  be very complex. We don't really want to do sql or SP way.
Right now we want to run this rule engine for daily data at specific time.
But we also want to use it for live data later on. so we want to develop
this rule engine as a library to use for different source( could be from
database,  or live data feed etc.,).

3.How many rules are you talking about? Could there be all kinds of rules,
or
just a fixed set of rules with different parameters?

- we are working on iterative model. now we have with 10 rules and develop
rules engine for those and keep adding new rules. We are expecting to have
1200 rules at the end.

I think we would want to define different rule template.  then user can
create different rules for the rule template with varying parameters.

Here are few rules we have right now.
 a.one set of rules will compare values from field combination. It it does
not match the expected value, then log it in the database with that rule
info. each rule in this category will check for different field combination.
 b. another set of rule  can be complicated. some examples are given below.
         I.To watch  for field combo for n number of days. i can create a
rule  saying that " if field1 =''a"    and field2 = 'b' combo occurs for 3
days continuously in daily data, then log those data into database.
         or
         II. i can also create a rule like this. If % variation of total
count of daily data for field combo field1 =''a" & field2 = 'b' between two
days > 10%, then log those data into database. Ex: 1 day  i get 100 data for
this combo and next day i get only 50 data for this combo. then I need to
log this data(1 day count & 2nd day count) into database.

Please guide me on the right approach. 






techluver wrote:
> 
> 
> 1.I would fist step back and ask the question if a rules engine is the
> right
> solution to your problem.
> 
> - Actually I do have same question and need experts help to found out
> right answer.
> 
> 1.Is this data you are getting in daily in a single table, multiple
> tables? 
> - we want to assume that it is from multiple tables.
> 
> 2.How complex is the data? Is the data simple enough that rules could just
> be
> some SQL statements/Stored procs?
> 
> - data won't be very complex. We don't really want to do sql or SP way.
> Right now we want to run this rule engine for daily data at specific time.
> But we also want to use it for live data later on. so we want to develop
> this rule engine as a library to use for different source( could be from
> database,  or live data feed etc.,). 
> 
> 3.How many rules are you talking about? Could there be all kinds of rules,
> or
> just a fixed set of rules with different parameters?
> 
> - we are working on iterative model. now we have with 10 rules and develop
> rules engine for those and keep adding new rules. We are expecting to have
> 1200 rules at the end.
> right now we have two sets of rules. 
>  a.one set of rules will come compare values from field combination. It it
> does not match the expected value, then log it in the database with that
> rule info. each rule in this category will check for different field
> combination. i think we can say this category as "fixed set of rules with
> different parameters"
>  b. another set of rule  can be complicated. some examples are given
> below.
>          I.To watch  for field combo for n number of days. i can create a
> rule  saying that " if field1 =''a"    and field2 = 'b' combo occurs for 3
> days continuously in daily data, then log those data into database.
>          or 
>          II. i can also create a rule like this. If % variation of total
> count of daily data for field combo field1 =''a" & field2 = 'b' between
> two days > 10%, then log those data into database. Ex: 1 day  i get 100
> data for this combo and next day i get only 50 data for this combo. then I
> need to log this data(1 day count & 2nd day count) into database. 
> 
> Please guide me on the right approach.
> 
> 
> 
> 
> dave sinclair wrote:
>> 
>> I would fist step back and ask the question if a rules engine is the
>> right
>> solution to your problem.
>> 
>> Is this data you are getting in daily in a single table, multiple tables?
>> How complex is the data? Is the data simple enough that rules could just
>> be
>> some SQL statements/Stored procs?
>> 
>> How many rules are you talking about? Could there be all kinds of rules,
>> or
>> just a fixed set of rules with different parameters?
>> 
>> On Fri, Oct 3, 2008 at 1:55 PM, techluver <techluver007 at gmail.com> wrote:
>> 
>>>
>>> Thanks David.
>>> I would like to give more details on my requirement. Please throw your
>>> ideas
>>> for good design using drools.
>>>
>>> 1.I'll get huge amount of data in database(60 or 70 millions) daily. I
>>> would
>>> like to apply the rules on those records and store the data which
>>> satisfied
>>> the rules. So my rule engine has to run daily at specific time.
>>>
>>> 2.I also have to provide the reporting GUI to business users to generate
>>> report based on rules
>>>
>>> 3. users also need user friendly GUI to manage the rules something like
>>> BRMS(this would be last thing user want)
>>>
>>> 4. till I develop rules manager gui, I should be able to manage the
>>> rules
>>> from the database.
>>>
>>> 5. I don''t need to load and apply the rules dynamically. But If I
>>> add/remove any  rule, then it should be loaded/removed during next run
>>> of
>>> rule engine.
>>>
>>> I'm not really sure whether to use drools' BRMS or not here. I think i
>>> need
>>> database driven rule engine.
>>>
>>> Please share your ideas for this requirement on what kind of design I
>>> need
>>> here.
>>>
>>>
>>> Thanks in advance.
>>>
>>>
>>>
>>>
>>> techluver wrote:
>>> >
>>> > I'm new to drools world. Kindly clarify following things
>>> >
>>> > My requirement is to log each data in the DB(that meets any of the
>>> > business rule) along with rule info.
>>> >
>>> > if I use BRMS ( i think it maintains the rules separately in its own
>>> db
>>> > table)
>>> >
>>> > How I can associate each log with the rule info?
>>> >
>>> > Approach 1. Can I access BRMS's rules directly and associate it in my
>>> > logging?
>>> >
>>> > Approach 2. I can't do Approach 1, then I'm thinking to have RULES
>>> table
>>> > in my app db.Then whenever any rule is added using BRMS, Then add an
>>> entry
>>> > into RULES table and use this table for my logging
>>> >
>>> > Please advise me on the right approach.
>>> >
>>> > Thanks in advance.
>>> >
>>>
>>> --
>>> View this message in context:
>>> http://www.nabble.com/BRMS-rule-info-logging-tp19766982p19802730.html
>>> Sent from the drools - user mailing list archive at Nabble.com.
>>>
>>> _______________________________________________
>>> rules-users mailing list
>>> rules-users at lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/rules-users
>>>
>> 
>> _______________________________________________
>> rules-users mailing list
>> rules-users at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/rules-users
>> 
>> 
> 
> 

-- 
View this message in context: http://www.nabble.com/BRMS-rule-info-logging-tp19766982p19842489.html
Sent from the drools - user mailing list archive at Nabble.com.




More information about the rules-users mailing list